CN115242721A - Embedded system and data flow load balancing method based on same - Google Patents
Embedded system and data flow load balancing method based on same Download PDFInfo
- Publication number
- CN115242721A CN115242721A CN202210783048.9A CN202210783048A CN115242721A CN 115242721 A CN115242721 A CN 115242721A CN 202210783048 A CN202210783048 A CN 202210783048A CN 115242721 A CN115242721 A CN 115242721A
- Authority
- CN
- China
- Prior art keywords
- application
- load balancing
- information
- node
- copy
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/10—Active monitoring, e.g. heartbeat, ping or trace-route
Abstract
The invention provides an embedded system and a data flow load balancing method based on the embedded system. The invention adopts a design method combining elastic expansion and load balancing, supports data stream load balancing on applications with complicated and changeable data stream topologies in an environment with limited computing power, mainly solves the problem of insufficient load balancing real-time performance in scenes with complicated and changeable data stream topologies, realizes an embedded system data stream load balancing framework, and supports intelligent radar information processing.
Description
Technical Field
The invention belongs to the technical field of communication, and relates to an embedded system and a data flow load balancing method based on the embedded system.
Background
The load balancing is based on reasonably distributing data, so that a fixed number of data streams are distributed to each processing device in a balanced manner for processing. In the application of the load balancing technology, data flow can be guided to be transferred in two links or a plurality of links, namely the problem of multi-link load balancing.
In an embedded environment, the challenges of data stream load balancing are mainly that there is no universal load balancing framework, the current commercial load balancing framework only performs corresponding data distribution processing for network data requests, and the load balancing framework is operated in a network forwarding gateway and is not suitable for an embedded platform application scenario. Data stream load balancing for applications with complex and variable data stream topologies in embedded systems with low cost, low energy consumption and limited computing power faces challenges.
Disclosure of Invention
In order to solve the problems existing in the prior art, the invention provides an embedded system, which comprises:
the node manager starts a back-end application executable program by calling the interface and sends node state information to the node controller;
the node controller receives the node state information sent by the node manager in real time, performs node addition and node removal management, collects CPU load information of application and sends the CPU load information to the copy controller;
the application manager receives the upper computer application deployment command, downloads the application and distributes the application operation related files to corresponding nodes;
the copy controller monitors the node state information and the CPU load information of the application, and executes an application elastic scaling algorithm to control the number of the applied copies;
the load balancing controller comprises a load balancing control module and a command sending module, the load balancing control module generates load balancing control information based on the application copy number and the application receiving and sending number theme, and the command sending module is responsible for sending the load balancing control information to the corresponding application copy;
the application and application copy comprises a command receiving module and a data flow control module, wherein the command receiving module is responsible for receiving control information from the load balancing controller, and the data flow control module is used for controlling the sending rhythm of the application based on the load balancing control information.
The method is realized based on the embedded system and specifically comprises the following steps:
the application manager distributes the application to the node managers on the corresponding nodes, and the node managers start the application to run and send node state information to the node controllers;
the node controller receives the node state information sent by the node manager in real time, collects the CPU load information of the application and sends the CPU load information to the copy controller;
the copy controller executes an application elastic scaling algorithm to control the number of the applied copies according to the node state information and the applied CPU load information;
the load balancing controller generates load balancing data flow control information based on the application copy number and the application receiving and sending number theme, and sends the data flow control information to the corresponding application copy;
the application copy controls the sending rhythm of the application based on the received data flow control information, thereby completing the load balancing processing process of the data flow;
and the load balancing controller periodically monitors whether the number of the application copies in the system is updated or not, and performs a load balancing control process.
Further, the elastic expansion control process specifically includes:
the copy controller monitors application CPU load information;
and judging whether the application CPU load information is greater than an upper limit threshold, if so, popping up an application copy and ending the process, otherwise, judging whether the application CPU load information is less than a lower limit threshold, if so, reducing the application copy and ending the process, and if not, directly ending the process.
Further, the load balancing control process specifically includes the following steps:
monitoring whether the number of the application copies is updated or not, and if not, directly ending the process;
if so, controlling the sending proportion of the application copies, sending control information to the corresponding application copies, judging whether the copy information is updated successfully, if so, directly ending the process, otherwise, returning to sending the control information to the corresponding application copies, and repeating the steps until the process is ended.
Further, the load balancing data flow control information includes number ratio example information. Compared with the prior art, the invention has the following beneficial effects:
the invention adopts a design method combining elastic expansion and load balancing to realize data stream load balancing on the application with complicated and changeable data stream topology in an embedded system with low cost, low energy consumption and limited computing capacity. The method supports a load balancing strategy in an environment with limited computing power, improves the data processing capability of the embedded platform and the utilization rate of computing resources and network resources, improves the reliability of the embedded platform, and meets the high-performance processing requirement in the radar field.
Drawings
Fig. 1 is a framework diagram of load balancing of data flows of an embedded system according to a first embodiment.
Fig. 2 is a flowchart of the elastic expansion control according to the second embodiment.
Fig. 3 is a flowchart of load balancing control according to the second embodiment.
Detailed Description
The present invention will be described in further detail with reference to the following examples and the accompanying drawings.
The first embodiment is as follows:
the embodiment provides an embedded system, as shown in fig. 1, including:
the node manager mainly comprises an application starting module and a heartbeat module. The application starting module starts a back-end application executable program through an operating system calling interface, and meanwhile, the node manager sends heartbeat information (OK: representing normal) to the node controller, wherein the sending beat is configurable and defaults to 1 second.
The node controller mainly comprises a node monitoring and management module and is responsible for receiving heartbeat information of the node manager in real time, when the node manager is abnormal, the node controller detects that the heartbeat information of the node manager is continuously lost for three times, and the node state is set to be abnormal. And performing node adding and node removing management through the node controller, collecting CPU load information of the application and sending the CPU load information to the replica controller.
The application manager mainly comprises an application distribution module which is responsible for receiving the deployment command of the application of the upper computer, downloading the application and distributing the application operation related files to the corresponding nodes.
The copy controller mainly comprises a monitoring module and an application copy control module, is responsible for monitoring node state information and applied CPU load information, and executes an application elastic expansion algorithm to control the copy number of the application.
The load balancing controller mainly comprises a load balancing control module and a command sending module, wherein the load balancing control module generates load balancing control information (such as the information of the number of the sender ratio) based on the number of the application copies and the theme of the number of the application transceivers, and the command sending module is responsible for sending the load balancing control information to the corresponding application copies.
The application and application copy mainly comprises a command receiving module and a data flow control module, wherein the command receiving module is responsible for receiving control information from the load balancing controller, and the data flow control module is used for controlling the sending rhythm of the application based on the sending ratio proportion.
Example two
The embodiment provides a method for implementing data stream load balancing, which is implemented based on an embedded system of the first embodiment and specifically includes the following steps:
downloading the application in the warehouse to a background by executing an application deployment command at a user, and generating a deployment scheme by a scheduler according to the resource requirement of the application;
and the application manager distributes the application to the corresponding node to run according to the deployment scheme. In the application running process, as the data volume processed by a task is larger and larger, an application program cannot meet the processing requirement, and at the moment, a node controller collects CPU load information of the application and sends the information to a copy controller;
the copy controller executes an application elastic scaling algorithm to control the number of the applied copies according to the node state information and the applied CPU load information;
the load balancing controller generates load balancing data flow control information (such as the ratio of the number of the data flow to the number of the data flow) based on the number of the application copies and the theme of the application sending and receiving number, and sends the data flow control information to the corresponding application copies;
the application copy controls the sending rhythm of the application based on the received data flow control information, thereby completing the load balancing processing process of the data flow;
the node controller collects CPU load information of the application and sends the CPU load information to the copy controller, and the copy controller reduces an application copy or pops up the application copy through an elastic expansion control process; as shown in fig. 2, the elastic expansion control process specifically includes:
the copy controller monitors application CPU load information;
judging whether the application CPU load information is larger than an upper limit threshold value, if so, popping up an application copy and ending the process, otherwise, judging whether the application CPU load information is smaller than a lower limit threshold value, if so, reducing the application copy and ending the process, and if not, directly ending the process;
meanwhile, the load balancing controller periodically monitors whether the number of the application copies in the system is updated, and performs a load balancing control process, as shown in fig. 3, where the load balancing control process specifically includes the following steps:
monitoring whether the number of the application copies is updated, if not, directly ending the process;
if yes, controlling the sending proportion of the application copies, sending control information to the corresponding application copies, judging whether the copy information is updated successfully, if yes, directly ending the process, otherwise, returning to sending the control information to the corresponding application copies, and repeating the steps until the process is ended.
The application copy (application a-application An in fig. 1) receives control information from the load balancing controller and controls the transmission rate rhythm of the application based on the transmission rate ratio.
The embodiment supports the load balancing implementation method of the hybrid embedded platform such as the X86, the DSP and the Feiteng, and improves the data processing capacity of the embedded platform and the utilization rate of computing resources and network resources.
The invention establishes an embedded system data flow load balancing framework, and realizes data flow load balancing on applications with complex and changeable data flow topology in an embedded system with low cost, low energy consumption and limited computing capacity; the invention adopts a design method combining elastic expansion and load balance, improves the reliability of the embedded platform and meets the high-performance processing requirement in the radar field; the invention provides a load balancing implementation method of a hybrid embedded platform such as an X86, DSP and Feiteng platform, and the like, which improves the data processing capacity of the embedded platform and the utilization rate of computing resources and network resources.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (5)
1. An embedded system and a data flow load balancing method based on the embedded system are characterized by comprising the following steps:
the node manager starts a back-end application executable program by calling the interface and sends node state information to the node controller;
the node controller receives the node state information sent by the node manager in real time, performs node addition and node removal management, collects the CPU load information of the application and sends the CPU load information to the copy controller;
the application manager receives the upper computer application deployment command, downloads the application and distributes the application operation related files to corresponding nodes;
the copy controller monitors the node state information and the CPU load information of the application, and executes an application elastic scaling algorithm to control the number of the applied copies;
the load balancing controller comprises a load balancing control module and a command sending module, the load balancing control module generates load balancing control information based on the application copy number and the application receiving and sending number theme, and the command sending module is responsible for sending the load balancing control information to the corresponding application copy;
the application and application copy comprises a command receiving module and a data flow control module, wherein the command receiving module is responsible for receiving control information from the load balancing controller, and the data flow control module is used for controlling the sending rhythm of the application based on the load balancing control information.
2. A data flow load balancing method, implemented based on the embedded system of claim 1, specifically comprising the steps of:
the application manager distributes the application to the node managers on the corresponding nodes, and the node managers start the application to run and send node state information to the node controllers;
the node controller receives the node state information sent by the node manager in real time, collects the CPU load information of the application and sends the CPU load information to the copy controller;
the copy controller executes an application elastic scaling algorithm to control the number of the applied copies according to the node state information and the applied CPU load information;
the load balancing controller generates load balancing data flow control information based on the application copy number and the application receiving and sending number theme, and sends the data flow control information to the corresponding application copy;
the application copy controls the sending rhythm of the application based on the received data flow control information, thereby completing the load balancing processing process of the data flow;
and the load balancing controller periodically monitors whether the number of the application copies in the system is updated or not, and performs a load balancing control process.
3. The data flow load balancing method according to claim 2, wherein the elastic scaling control flow specifically includes:
the copy controller monitors application CPU load information;
judging whether the application CPU load information is larger than an upper limit threshold value, if so, popping up an application copy and ending the process, if not, judging whether the application CPU load information is smaller than a lower limit threshold value, if so, reducing the application copy and ending the process, and if not, directly ending the process.
4. The data flow load balancing method according to claim 3, wherein the load balancing control flow specifically includes the following steps:
monitoring whether the number of the application copies is updated or not, and if not, directly ending the process;
if so, controlling the sending proportion of the application copies, sending control information to the corresponding application copies, judging whether the copy information is updated successfully, if so, directly ending the process, otherwise, returning to sending the control information to the corresponding application copies, and repeating the steps until the process is ended.
5. The data flow load balancing method of claim 4, wherein the load balancing data flow control information comprises a number ratio example information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210783048.9A CN115242721A (en) | 2022-07-05 | 2022-07-05 | Embedded system and data flow load balancing method based on same |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210783048.9A CN115242721A (en) | 2022-07-05 | 2022-07-05 | Embedded system and data flow load balancing method based on same |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115242721A true CN115242721A (en) | 2022-10-25 |
Family
ID=83671100
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210783048.9A Pending CN115242721A (en) | 2022-07-05 | 2022-07-05 | Embedded system and data flow load balancing method based on same |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115242721A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100228819A1 (en) * | 2009-03-05 | 2010-09-09 | Yottaa Inc | System and method for performance acceleration, data protection, disaster recovery and on-demand scaling of computer applications |
| CN102244685A (en) * | 2011-08-11 | 2011-11-16 | 中国科学院软件研究所 | Distributed type dynamic cache expanding method and system supporting load balancing |
| CN107018197A (en) * | 2017-04-13 | 2017-08-04 | 南京大学 | A kind of holding load dynamic retractility mobile awareness Complex event processing method in a balanced way |
| CN114389955A (en) * | 2022-03-02 | 2022-04-22 | 中国电子科技集团公司第十四研究所 | Embedded platform heterogeneous resource pooling management method |
| CN114598591A (en) * | 2022-03-07 | 2022-06-07 | 中国电子科技集团公司第十四研究所 | Embedded platform node fault recovery system and method |
-
2022
- 2022-07-05 CN CN202210783048.9A patent/CN115242721A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100228819A1 (en) * | 2009-03-05 | 2010-09-09 | Yottaa Inc | System and method for performance acceleration, data protection, disaster recovery and on-demand scaling of computer applications |
| CN102244685A (en) * | 2011-08-11 | 2011-11-16 | 中国科学院软件研究所 | Distributed type dynamic cache expanding method and system supporting load balancing |
| CN107018197A (en) * | 2017-04-13 | 2017-08-04 | 南京大学 | A kind of holding load dynamic retractility mobile awareness Complex event processing method in a balanced way |
| CN114389955A (en) * | 2022-03-02 | 2022-04-22 | 中国电子科技集团公司第十四研究所 | Embedded platform heterogeneous resource pooling management method |
| CN114598591A (en) * | 2022-03-07 | 2022-06-07 | 中国电子科技集团公司第十四研究所 | Embedded platform node fault recovery system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107087019B (en) | Task scheduling method and device based on end cloud cooperative computing architecture | |
| CN111131401B (en) | Service request response method and system | |
| CN109756912B (en) | Multi-user multi-base station joint task unloading and resource allocation method | |
| CN112650581A (en) | Cloud-side cooperative task scheduling method for intelligent building | |
| CN113645146B (en) | New stream density-based software defined network controller load balancing method and system | |
| CN102902594A (en) | Resource management system and resource management method | |
| CN112416603A (en) | Combined optimization system and method based on fog calculation | |
| Zhao et al. | PLOFR: An online flow route framework for power saving and load balance in SDN | |
| CN111245630A (en) | Method for simultaneously upgrading multiple devices in local area network | |
| CN105207856A (en) | Load balancing system and method based on SDN virtual switch | |
| CN115242721A (en) | Embedded system and data flow load balancing method based on same | |
| CN110850957B (en) | Scheduling method for reducing system power consumption through dormancy in edge computing scene | |
| CN116614517B (en) | Container mirror image preheating and distributing method for edge computing scene | |
| CN107948330A (en) | Load balancing based on dynamic priority under a kind of cloud environment | |
| CN116436923A (en) | Method and system for autonomously generating and matching on-demand networking service instance | |
| CN113822485B (en) | Power distribution network scheduling task optimization method and system | |
| CN115550438A (en) | Internet of things message processing method, device, system, equipment and storage medium | |
| CN111770515B (en) | SDN controller deployment method based on network cost optimization | |
| CN114760015B (en) | EMS remote adjustment remote control success rate improving method based on redundant design and strategy control | |
| CN113915741A (en) | Instruction sending method, device and system | |
| CN106899392B (en) | Method for carrying out fault tolerance on instantaneous fault in EtherCAT message transmission process | |
| CN113722096B (en) | CPU frequency adjusting method, system, equipment and storage medium of edge computing node | |
| CN112311584B (en) | Deployment method of software defined network controller | |
| CN117076057B (en) | AI service request scheduling method, device, equipment and medium | |
| Kobari et al. | A reinforcement learning based edge cloud collaboration |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |